Diffeomorphic Spectral Matching
This matlab code finds a diffeomorphic (smooth and bijective) correspondence map between cortical surfaces and generic meshes.It is based on the paper Diffeomorphic Spectral Matching, IPMI 2013
Source code (Matlab)
This is preliminary code with little comments,
however, cleaned up code should replace it soon, check regularly
Spectral Demons
This is Matlab code for registering two images which have very large deformations.It is based on the paper Spectral Demons, ECCV 2012
Brief Explanation (movie 30sec or 7min)
Source code (Matlab)
This is preliminary code with little comments,
however, cleaned up code should replace it soon, check regularly
FOCUSR for Spectral Matching
This matlab code implements FOCUSR, an algorithm for finding mesh correspondences using spectral matching.It is based on the paper FOCUSR, PAMI 2013, and IPMI 2011
Additional Page
Source code (Matlab)
This is preliminary code with little comments,
however, cleaned up code should replace it soon, check regularly
libvd
This is a very simple OpenGL visualization library. It has a volume renderer. The main purpose of this library is to visualize medical volumes. As the code is simple and well isolated, it can easily be reused in other projects. It is just perfect for learning or great if you like to h ack your own code.
VolumeDev
This project is meant for quick algorithm development involving medical volume visualization. Working on an algorithm only requires implementation of its core, visualization and gui are all provided. If not, new functionalities can be added through simple plugins.
The project is currently in a very early stage.
Source code should be available soon.
Deformation with Thin Plate Splines
By using a set of control points to define where some parts of the image will move, Thin Plate Splines are used to interpolate the deformed image. This manipulation is convenient for non-rigid registration.
Registration with mutual information
By doing this project, my partner and I aimed to learn the basics of registration. We compared different voxel similarity measures such as cross-correlation, mutual information, and mutual information combined with gradient information.